Summary of Reuse Out-of-year Data to Enhance Land Cover Mapping Via Feature Disentanglement and Contrastive Learning, by Cassio F. Dantas (umr Tetis et al.
Reuse out-of-year data to enhance land cover mapping via feature disentanglement and contrastive learning
by Cassio F. Dantas, Raffaele Gaetano, Claudia Paris, Dino Ienco
First submitted to arxiv on: 17 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel deep learning framework, REFeD, is proposed to tackle the challenge of making value of historical land use/land cover (LULC) data for agricultural territory management and environmental monitoring. By leveraging domain adaptation and generalization techniques, REFeD combines remote sensing and reference data from two different domains to enhance the current LULC mapping process. The framework uses contrastive learning-based disentanglement strategies to recover intrinsic information related to the downstream LULC mapping task, alleviating distribution shifts between domains. Additionally, an effective supervision scheme is employed to enforce feature disentanglement at multiple levels. Experimental results on two study areas demonstrate the quality of REFeD and show that out-of-year data from the same or similar study sites can be a valuable additional source of information for LULC mapping. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A team of researchers has come up with a way to use old land use/land cover maps to make new ones. This is important because it can help farmers and scientists make better decisions about how to grow food and protect the environment. The team used special computer algorithms to combine old and new data from different places, which helps reduce the need for expensive and time-consuming field campaigns. They tested their approach on two areas with very different landscapes and found that it worked well. |
Keywords
» Artificial intelligence » Deep learning » Domain adaptation » Generalization
